Genome-Wide Identification of the 1-Aminocyclopropane-1-carboxylic Acid Synthase (ACS) Genes and Their Possible Role in Sand Pear (Pyrus pyrifolia) Fruit Ripening

Ethylene production is negatively associated with storage life in sand pear (Pyrus pyrifoliaNakai), particularly at the time of fruit harvest. 1-Aminocyclopropane-1-carboxylic acid synthase (ACS) is the rate-limiting enzyme in ethylene biosynthesisand is considered to be important for fruit storage life. However, the candidate ACS genes and their roles in sand pear remain unclear. The present study identified 13ACS genes from the sand pear genome.Phylogenetic analysiscategorizedthese ACS genesinto four subgroups (type Ⅰ, type Ⅱ, type Ⅲ and putative AAT), and indicated a close relationship between sand pear and Chinese white pear (P. bretschneideri). According to the RNA-seq data and qRT-PCR analysis, PpyACS1, PpyACS2, PpyACS3, PpyACS8, PpyACS9, PpyACS12 and PpyACS13 were differently expressed in climacteric and non-climacteric-typepear fruits,‘Ninomiyahakuri’ and ‘Eli No.2′, respectively, during fruit ripening. In addition, the expressions of PpyACS2, PpyACS8, PpyACS12 and PpyACS13 werefound to be associated with system 1 of ethylene production, while PpyACS1, PpyACS3, and PpyACS9 werefound to be associated with system 2, indicating that these ACS genes have different roles in ethylenebiosynthesis during fruit development. Overall, our study provides fundamental knowledge onthe characteristics of the ACS gene family in sand pear, in addition to their possible roles infruit ripening.


Introduction
Ethylene is a phytohormone that participates in many events of plant growth and development, like organ abscission, stress responses, fruit ripening and senescence [1,2]. Generally, ethylene biosynthesis is divided into two systems, named system 1 and system 2. System 1allows basal ethylene production in vegetative tissues and unripe fruit. System 2 involves the burst of ethylene production during the ripening of climacteric fruit. Moreover, system 1 is regulated in an auto-inhibitory manner, while system 2 is regulated in an autocatalytic manner. Fruit ripening behavior is generally characterized as being of a climacteric or a non-climacteric type, depending on whether or not the process is associated with a peak of ethylene production and a rise in respiration rate [3][4][5].
Ethylene is synthesized from S-adenosyl-L-methionine (SAM) via 1-aminocyclopropane-1-carboxylic acid (ACC). The two major processes in ethylene biosynthesis are the metabolic conversion of SAM to ACC by ACC synthase (EC 4.4.1.14) and secondly the conversion of ACC to ethylene by ACC oxidase (EC 1.14.17.4) [6]. In plant species, ACS is the ratelimiting enzyme in ethylene synthesis and is encoded by a widely diverse multi-gene family [7]. Ethylene biosynthesis is a complicated process involving numerous ACS genes.
The variable expression of ACS genes may be implicated in the synthesis of ethylene at a specific developmental stage. For example, there are 14 ACS genes in tomato (Solanum lycopersicum), among them SlACS1A, SlACS2, SlACS4, SlACS11 and SlACS12 showed different expression profiles during fruit ripening [8]. Three ACS genes (MdACS1, MdACS3a, and MdACS6) in apple have been extensively examined, and each of them was found to express distinctly [9][10][11]. MdACS3a was found to be expressed abundantly one month before fruit ripening, and its alleles are crucial in controlling fruit shelf-life, whereas MdACS1 was shown to be responsible for the burst of ethylene production in system 2. MdACS6 was expressed in the early stages of fruit development, before MdACS3a and MdACS1 expression began [11].Among six Pp-ACSs isolated from the peach genome, Pp-ACS1 and Pp-ACS4 showed ripening-related increased expression during fruit development and ripening [12].
The sand pear (Pyrus pyrifolia Nakai) is widely distributed in southern China, where the ancestral pear species originated, and where a large collection of germplasm, including wild genotypes, landraces, and improved cultivars, exists [13,14]. Sand pear exhibits both climacteric and non-climacteric behavior depending on the cultivar or genotype, making it a suitable material for studying the fundamental basis of the various ripening behaviors [4]. In this study, the genome-wide analyses of ACS genes were conducted across five Rosaceae and two non-Rosaceae species. A total of 13 PpyACSs were identified in sand pear fruit and RNA-seq analysis showed the great variations between their expression levels. To further understand the role of ACS in fruit ripening, the predominantly expressed six ACS genes were selected and their expression profiles were compared in the fruits 'Ninomiyahakuri' and 'Eli No. 2' at different ripening stages. The findings of this study could help researchers better understand the molecular mechanism of fruit ripening in sand pear.

Plant Materials and Treatments
Two sand pear cultivars ('Ninomiyahakuri' and 'Eli No. 2') grown in Wuhan City, Hubei Province, P.R. China, were used in this study ( Figure S1). Interestingly, both cultivars have different ripening behavior. Comparatively, fruits of 'Ninomiyahakuri' yield more ethylene during fruit ripening, whereas the fruits of 'Eli No. 2' yield less ethylene. The fruits of 'Ninomiyahakuri' and 'Eli No. 2' were collected 30 days before harvest (30 DBH; commercial maturity = 20 July 2017) and were kept at room temperature (24 • C) for 20 days. For lab analysis, fruits were sampled after an interval of 5 days after harvest (1,5,10,15,20 DAH). Five fruits were sampled at each sampling point for determination of ethylene production. Then the fruits were sliced, frozen in liquid nitrogen and stored at −80 • C for RNA extraction.
For spatial gene expression analysis, young fruit, mature fruit, flowers, leaf and stem were collected from 'Eli No. 2' pear trees. All samples were processed and stored as discussed above, for further analysis.

Measurements of Ethylene Production
The ethylene production was estimated with a gas chromatograph (Agilent 7890A, Santa Clara, CA, USA) fitted with a flame ionization detector. Briefly, fruits were placed in an airtight container (2.0 dm 3 ) equipped with septa and maintained at 24 • C for 1 h. Afterwards, 1 cm 3 of gas was sampled through the headspace of the container by a syringe.

RNA Isolation and RNA-Seq
Total RNA was extracted with TRIzol reagent (Invitrogen, Carlsbad, CA, USA) following the manufacturer's instructions. RNA concentration was measured by NanoDrop 2000 (Thermo, Waltham, MA, USA) and RNA integrity was evaluated by the RNA Nano 6000 Assay Kit using the Agilent Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA, USA).
Transcriptome libraries were constructed from samples of 'Ninomiyahakuri' and 'Eli No. 2' collected at three ripening stages: 1 (pre-ripening, 30 days before harvest), 2 (at harvest) and 3 (fruit senescence, 10 days after harvest). 1 µg RNA per sample was used for RNA-seq, which was carried out by Beijing Biomarker Technology Co., Ltd. (Beijing, China). In short, RNA-seq libraries were constructed using NEBNext ® Ultra™ RNA Library Prep Kit for Illumina ® (NEB, Ipswich, MA, USA) according to the manufacturer's protocol [15]. The cDNA library samples were sequenced by the Illumina Hi-Seq 2000 Sequencer. The transcriptome was aligned to the sand pear reference genome using Top Hat and the read count for each gene was obtained by the Cufflinks software. Quantification of gene expression levels were measured in fragments per kilobase of transcript per million fragments mapped (FPKM), and the DEGSeq R package (1.10.1) was used to analyze the differential expression of the pairwise comparisons [16].

Genome-Wide Identification of ACS Genes
In order to recognize the ACS gene family in P. pyrifolia and other four Rosaceae species (P. bretschneideri, P. communis, Malus domestica and Prunus persica), we searched the Genome Database for Rosaceae by BLASTP and BLASTN, using AtACSs sequences as queries [17,18]. The domain composition of ACSs were predicted by SMART (http://smart. embl-heidelberg.de/, accessed on 10 March 2021) and Pfam (http://pfam.san ger.ac.uk/, accessed on 10 March 2021). All of the identified ACS genes were analyzed using the NCBI's conserved domain database (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi, accessed on 12 March 2021) to determine the domains they possess. Multiple sequence alignment of ACS proteins was performed with Clustal X, the redundant or incomplete sequences were then removed. Additionally, ACS genes were named according to the guideline developed for Rosaceae family members [19].
The amino acid length and chromosome information were obtained from the genomic file. Protein molecular weights (Da) and isoelectric points (pI) were calculated by the online EXPASY server (http://web.expasy.org/protparam/, accessed on 12 March 2021). Gene motif analysis was performed with MEME v5.0.4 (http://meme-suite.org/tools/ meme, accessed on 13 March 2021). Gene structure, motifs and chromosome location were presented using TBtools [20]. The Plant-mPLoc (http://www.csbio.sjtu.edu.cn/cgibin/ PlantmPLoc.cgi, accessed on 15 March 2021) was performed to predict the subcellular localization of the ACS proteins. Cis-acting regulatory elements in the 2.0 kb of 5 -UTR of PpyACS were identified by the Plant-CARE database [21].

Phylogenetic Analysis
To identify orthologous genes, OrthoVenn2 was used to search against seven species genomes (Supplementary Table S1) using PpyACS as queries [22,23]. ACS amino acid sequences were aligned by using the MUSCLE program, and phylogenetic trees were built with the neighbor-joining method and 1000 bootstrap iterations in MEGAv7.0 software [24].

Quantitative Real-Time PCR (qRT-PCR) Analysis
The primers were designed by the Primer 6.0 program and listed in Table S2. Synthesis of the first-strand cDNA was conducted with PrimeScript™ RT reagent Kit with gDNA Eraser (TaKaRa, Shiga, Japan). cDNA was diluted 2-fold, and used as a template for qPCR. The PCR mix (50 mm 3 ) contained: 1.0 mm 3 Ex Taq and 25.0 mm 3 2X Ex Taq Buffer (both Takara), 1.0 mm 3 dNTP Mix, 1.5 mm 3 of each primer (10X), 5.0 mm 3 cDNA, and 15.0 mm 3 PCR-Grade Water. The melt curve was assessed from 65 to 95 • C, with 0.5 • C increments. Relative levels of gene expression were calculated with the 2 −∆∆Ct method [25]. qRT-PCR gene expression was analyzed by CFX Connect™ (BIO-RAD, Hercules, CA, USA) using the PP2A gene as an internal control [26]. All analyses were achieved by three independent biological replicates.

ACS Gene Sequence Identification
In this study, 56 ACS genes from pear were identified and it included 13 genes from apple and 8 from peach, 10 genes from Chinese white pear, 12 genes from European pear as well as 13 genes from sand pear. The detailed information of gene IDs, genomic positions, coding region lengths and translated protein sequence about pear ACS genes were performed in Table S3. Lineage-specific whole-genome duplication (WGD) was found to have occurred in the ancestor of pear and apple [27], which may have resulted in almost double the number of ACS genes in these four species than in peach. The numbers of ACS genes identified in three species of Pyrus spp. have slight differences that may be due to their genome assembly quality.
Based on the presence or absence of C-terminal phosphorylation motifs, ACS proteins are classified into three categories (type I, type II, and type III) [28]. Type I ACSs have a C-terminal segment with phosphorylation sites for both mitogen-activated protein kinases (MAPKs) and calcium-dependent protein kinases (CDPKs) (CDPKs). Type II isozymes only have the CDPK target site, but type III isozymes lack both the MAPK and the CDPK sites. There is also another ACS-like homolog, AtACS10 and AtACS12, which are thought to be amino acid transferases lacking ACS activity and are known as putative AAT [29].
All identified ACS genes contained a "1-Aminocyclopropane-1-carboxylic acid synthase" domain but belong to the different domain families: PLN02450 superfamily or PLN02450 exist in all types of ACS except type III, which specifically possess PLN2607. Interestingly, two type I ACS homologous genes (AtACS1 and AtACS2) in Arabidopsis have a PLN02376 domain, which is not present in other investigated plants (Table S3).
In sand pear, all PprACSs contained the seven conserved regions of ACS ( Figure 1) [23] and were located in 8 chromosomes unevenly: chromosomes 15 and 2 resided three genes, chromosomes 1 had two genes, and chromosomes 4, 6, 7, 8 and 14 only showed one gene ( Figure S2). Subcellular prediction results showed that all PpyACS proteins were localized in the chloroplast, except for type II ACS, which were cytoplasm-localized proteins. The length of PpyACS-encoded protein sequences ranged from 445 (PpyACS11) to 611 (PpyACS13) amino acids. The predicted molecular weights and pI ranged from 49.84 kDa (PpyACS11) to 67.55 kDa (PpyACS13) and from 5.63 (PpyACS10) to 8.76 (PpyACS12), respectively (Table 1). The exon-intron organizations and motifs of all PpyACS genes were examined in P. pyrifolia. As shown in Figure 2, 16 conserved motif sequences were detected (Table S4). All PpyACS contained motif 1~10. Motif 11 and 12 were only found in type I ACS, whereas motif 15 was unique to putative AAT genes. Motif 13 was found in type I and type III ACS. In four type II ACS genes, motif 14 was present in PpyACS1 and PpyACS5, whereas motif 16 only existed in PpyACS1 and PpyACS5. In general, the motifs are quite conserved within each type. In addition, the presence of a varying number and length of introns among the PpyACSs contribute to significant variations in gene length. The number of introns per PpyACSs also varied from two to five: the type III PpyACSs were characterized by two introns and the other PpyACSs displayed three to four introns except for PpyACS13, which contained five introns.  The exon-intron organizations and motifs of all PpyACS genes were examined in P. pyrifolia. As shown in Figure 2, 16 conserved motif sequences were detected (Table S4). AllPpyACScontained motif 1~10. Motif 11 and 12 were only found intypeⅠ ACS, whereas motif 15 was unique toputative AAT genes. Motif 13 was found intypeⅠ and type Ⅲ ACS. In four typeⅡ ACS genes, motif 14 was present in PpyACS1 and PpyACS5, whereas motif 16 only existed in PpyACS1 and PpyACS5. In general, the motifs are quite conserved within each type. In addition, the presence of a varying number and length of introns among the PpyACSscontribute to significant variations in gene length. The number of introns per PpyACSs also varied from two to five: the type ⅢPpyACSs were characterized by two introns and the other PpyACSs displayed three to four introns except for PpyACS13, which contained five introns.  Table S4.
To further understand the transcriptional regulation and potential functions of PpyACS genes, the upstream promoter sequences of PpyACSs (2000 bp upstream of the start codon) were isolated and predicted (Figure 3).A number of cis-elements implicated  Table S4.
To further understand the transcriptional regulation and potential functions of PpyACS genes, the upstream promoter sequences of PpyACSs (2000 bp upstream of the start codon) were isolated and predicted ( Figure 3). A number of cis-elements implicated in hormone and stress responses were found in the upstream sequences of these PpyACSs. MeJA responsiveness elements (TGACG-and CGTCA-motif) were present in all PpyACSs except PpyACS13. ABRE, ABA-responsive element was identified in most of the PpyACSs, except PpyACS4, PpyACS11 and PpyACS12. Three elements involved in gibberellin response, namely GARE-motif, TATC-box and P-box, were identified in all PpyACSs, except PpyACS7. Auxin-responsive elements (AuxRR-core and TGA-element) were found in most of the PpyACSs, except PpyACS6, PpyACS7, PpyACS8 and PpyACS11. We also noticed that cis-elements associated with abiotic stress response were widely distributed within the promoters of PpyACSs. For example, an element essential for the anaerobic induction, named ARE, was identified in the PpyACSs, except PpyACS6 and PpyACS12. A low temperature-responsive element (LTR) was identified in PpyACS1, PpyACS3, PpyACS9 and PpyACS10, whereas TC-rich repeats involved in defense response were observed in PpyACS1 and PpyACS2. Meanwhile, a MYB transcription factor binding site (MBS) involved in drought stress response was present in four members, including PpyACS2, PpyACS7, PpyACS10 and PpyACS11. In addition, a wound-responsive element (WUN motif) was found in PpyACS5, PpyACS9 and PpyACS11.

Phylogenetic Analysis of Putative ACS Genes
A total of 88 ACS proteins from seven species were phylogenetically categorized into four subgroups that perfectly fit their types (Figure 4). Type II included the highest number of ACS genes (27), followed by type I, which contained 25 ACS genes. This result provides further evidence that type I and type II contains a greater percentage of ACS genes to other types. Interestingly, each species has two members belonging to putative AAT that were presumed without ACS activity. In addition, from species phylogeny based on ACS orthologs genes ( Figure S3, Table S5), a close relationship of ACS genes was observed between Pyrus and Malus, and the closest relationship was detected between P. pyrifolia and P. bretschneideri genes.

Phylogenetic Analysis of Putative ACS Genes
A total of 88 ACS proteins from seven species were phylogenetically categorized into four subgroups that perfectly fit their types (Figure 4). Type II included the highest num-

Expression of PpyACSs in Different Organs
The qRT-PCR analysis showed that PpyACS genes were differentially expre different tissues. Among 13 PpyACS genes, PpyACS1, PpyACS2,PpyACS3 and Pp were expressed highly in mature fruits. The other nine PpyACS genes were expre the remaining studied tissues/organs, such as PpyACS4 and PpyACS5, which were expressed in the leaf, while PpyACS5, PpyACS7, PpyACS8,PpyACS10, PpyACS PpyACS13 in flowers ( Figure 5).

Expression of PpyACSs in Different Organs
The qRT-PCR analysis showed that PpyACS genes were differentially expressed in different tissues. Among 13 PpyACS genes, PpyACS1, PpyACS2, PpyACS3 and PpyACS9 were expressed highly in mature fruits. The other nine PpyACS genes were expressed in the remaining studied tissues/organs, such as PpyACS4 and PpyACS5, which were highly expressed in the leaf, while PpyACS5, PpyACS7, PpyACS8, PpyACS10, PpyACS12 and PpyACS13 in flowers ( Figure 5).

RNA-Seq of Sand Pear Fruit and Identification of PpyACSs
In this study, six transcriptome libraries were constructed and sequenced for cv. 'Ninomiyahakuri' and 'Eli No. 2' in three ripening stages, including 1 (pre-ripening, 30 days before harvest), 2 (at harvest) and 3 (fruit senescence, 10 days after harvest). Over 89.2% of clean reads were uniquely mapped to the sand pear genome [30]. Expression of 13 PpyACSs in the pear genome was detected. The transcript abundance of PpyACSs varied greatly among members in different climacteric type sand pears, with that of one gene, PpyACS1, accounting for the most of total abundance in 'Ninomiyahakuri' (climacteric type) but none in 'Eli No. 2' (non-climacteric type). And six genes (PpyACS2, PpyACS3, PpyACS8, PpyACS9, PpyACS12and PpyACS13) were expressed during the fruit maturation in both cultivars but in different FPKM (Fragments per kilobase of transcript per million fragments mapped) ( Figure 6). Based on these results, seven genes were selected for further analysis.

RNA-Seq of Sand Pear Fruit and Identification of PpyACSs
In this study, six transcriptome libraries were constructed and sequenced for cv. 'Ninomiyahakuri' and 'Eli No. 2' in three ripening stages, including 1 (pre-ripening, 30 days before harvest), 2 (at harvest) and 3 (fruit senescence, 10 days after harvest). Over 89.2% of clean reads were uniquely mapped to the sand pear genome [30]. Expression of 13 PpyACSs in the pear genome was detected. The transcript abundance of PpyACSs varied greatly among members in different climacteric type sand pears, with that of one gene, PpyACS1, accounting for the most of total abundance in 'Ninomiyahakuri' (climacteric type) but none in 'Eli No. 2' (non-climacteric type). And six genes (PpyACS2, PpyACS3, PpyACS8, PpyACS9, PpyACS12 and PpyACS13) were expressed during the fruit maturation in both cultivars but in different FPKM (Fragments per kilobase of transcript per million fragments mapped) ( Figure 6). Based on these results, seven genes were selected for further analysis.  (Table S6).

Changes in Internal Ethylene Concentration in Flesh during Fruit Ripening
To further elucidate the relationship between the expression pattern of the seven PpyACS genes and internal ethylene concentration, dynamic changes in internal ethylene concentration in the fruits of two cultivars were determined. As shown in Figure 7, there was a rapid increase in internal ethylene concentration and a large amount of ethylene was produced in 'Ninomiyahakuri' fruits that showed ripening behavior typical to climacteric fruits. In contrast, almost undetectable ethylene production was observed in 'Eli No. 2' fruits during postharvest.   (Table S6).

Changes in Internal Ethylene Concentration in Flesh during Fruit Ripening
To further elucidate the relationship between the expression pattern of the seven PpyACS genes and internal ethylene concentration, dynamic changes in internal ethylene concentration in the fruits of two cultivars were determined. As shown in Figure 7, there was a rapid increase in internal ethylene concentration and a large amount of ethylene was produced in 'Ninomiyahakuri' fruits that showed ripening behavior typical to climacteric fruits. In contrast, almost undetectable ethylene production was observed in 'Eli No. 2' fruits during postharvest.  (Table S6).

Changes in Internal Ethylene Concentration in Flesh during Fruit Ripening
To further elucidate the relationship between the expression pattern of the seven PpyACS genes and internal ethylene concentration, dynamic changes in internal ethylene concentration in the fruits of two cultivars were determined. As shown in Figure 7, there was a rapid increase in internal ethylene concentration and a large amount of ethylene was produced in 'Ninomiyahakuri' fruits that showed ripening behavior typical to climacteric fruits. In contrast, almost undetectable ethylene production was observed in 'Eli No. 2' fruits during postharvest.

Expression of PpyACSs during Fruit Ripening
To validate expression profiles of the ACS genes isolated from transcriptome analysis, the seven PpyACSs were selected for qRT-PCR testing in fruits of cv. 'Ninomiyahakuri' and 'Eli No. 2' (Figure 8). ACS1 was dramatically induced by ripening in 'Ninomiyahakuri' and showed the highest expression during postharvest, but was undetectable throughout ripening in 'Eli No. 2', which means PpyACS1 is the dominant gene that regulates ripeningassociated ethylene production in climacteric fruit. The expression of PpyACS2 was also strongly increased during storage, not only in climacteric fruits but also in non-climacteric varieties, although the peak was approximately 2-fold lower than that in climacteric fruits. It is worth noting that three out of four type I genes have shown distinct expression patterns in fruit ripening, such as PpyACS3 and PpyACS9, which were highly expressed after harvest in climacteric fruits, but the expression level of PpyACS8 was higher at the pre-harvest stage than at the post-harvest stage in both climacteric and non-climacteric fruits. During storage, the expression level of PpyACS8 decreased first and then increased. Interestingly, PpyACS12 and PpyACS13, two ACS homologues of AtACS10 and AtACS12 which are presumed as amino acid transferases without ACS activity, were highly expressed at 30 days before commercial harvest and the expression level decreased slightly after harvest.

Expression of PpyACSs during Fruit Ripening
To validate expression profiles of the ACS genes isolated from transcriptome analysis, the seven PpyACSs were selected for qRT-PCR testing in fruits of cv.'Ninomiyahakuri' and 'Eli No. 2' (Figure 8). ACS1 was dramatically induced by ripening in 'Ninomiyahakuri' and showed the highest expression during postharvest, but was undetectable throughout ripening in 'Eli No. 2', which means PpyACS1 is the dominant gene that regulates ripening-associated ethylene production in climacteric fruit. The expression of PpyACS2 was also strongly increased during storage, not only in climacteric fruits but also in non-climacteric varieties, although the peak was approximately 2-fold lower than that in climacteric fruits. It is worth noting that three out of four type I genes have shown distinct expression patterns in fruit ripening, such as PpyACS3 and PpyACS9, which were highly expressed after harvest in climacteric fruits, but the expression level of PpyACS8 was higher at the pre-harvest stage than at the post-harvest stage in both climacteric and non-climacteric fruits. During storage, the expression level of PpyACS8 decreased first and then increased. Interestingly, PpyACS12 and PpyACS13,two ACS homologues of AtACS10 and AtACS12 which are presumed as amino acid transferases without ACS activity, were highly expressed at 30 days before commercial harvest and the expression level decreased slightly after harvest.

Discussion
ACS is encoded by a multigene family with unique spatial and temporal expression patterns [8,12,23,[31][32][33]. Although two ACSs have been found in sand pear [4], there are likely to be unidentified ACS genes in the pear genome, and their relationship with ripening behavior is unknown. Currently, the full-length genome of sand pear (P. pyrifolia) has been sequenced, providing valuable gene resources for further functional genomics study [30]. In this study, a total of 13 ACS genes were found in the genome of sand pear, and the number is the same as that identified in P. ussuriensis [34]. A comprehensive overview of the ACS gene family was undertaken, including the gene structures, conserved motifs, chromosome locations, phylogeny, and cis-elements in the promoter sequence, as well as expression patterns.
It was noticed that the number of species in which the fruits of different varieties exhibit both climacteric and non-climacteric behavior and the classification of fruits into climacteric and non-climacteric categories are an over-simplification [35]. Consistent with previous reports [4,36], substantial differences in ethylene production during fruit ripening was observed between two sand pear cultivars, which indicated that sand pear shows climacteric as well as non-climacteric behavior, depending on the genotype. Previous research showed that ethylene production in climacteric sand pears is attributed to PpyACS1 and PpyACS2 expression during the ripening stage [4], but the mechanism of fruit ripening in non-climacteric fruit is unclear.
In this study, genome-wide expression analysis based on RNA-seq data and qRT-PCR was carried out with seven genes identified as the most strongly expressed genes during fruit ripening, PpyACS1, PpyACS2, PpyACS3, PpyACS8, PpyACS9, PpyACS12 and PpyACS13 ( Figure 6, Table 2). PpyACS1 was pre-dominantly expressed only in climacteric fruits after harvest and in accordance with the change in system 2 ripening-associated ethylene production, which indicated that PpyACS1 may play key roles in the determination of ethylene production of climacteric fruit ripening. The same result was verified in previous studies [4,31,37,38]. Otherwise, as a type II gene, PpyACS1 was probably under the regulation of ETO1 (ETHYLENE OVERPRODUCER 1) [28], which needs further research. Type III ACS proteins lack both MAPK and CDPK sites for post-translational regulation and may be of minor importance in fruit ripening [12]. However, a type IIIACS gene, PPACS2 (PpyACS2), appears to be only expressed in the ripening of sand pear cultivars producing high or moderate levels of ethylene [4]. Surprisingly, our results showed that PpyACS2 was expressed highly, even in non-climacteric fruit after harvest, which indicated that PpyACS2 was responsible for system 1 ethylene biosynthesis in fruit ripening, similar to its ortholog gene, MdACS3, in apple.
Several investigations have revealed that type I functional divergence was the prevailing trend in the evolution of ACS subclades in plants [39]. In tomato fruit, it is well established that SlACS6 and SlACS1A expression is associated with system 1 of ethylene production, while SlACS4 and SlACS2 are associated with system 2 [8,40]. Interestingly, all these four ACSs belong to type I (Figure 4) [41]. However, no type I ACS genes were expressed in apple fruit [33]. In the present study, three type I genes were newly identified-PpyACS3, PpyACS8 and PpyACS9-and were involved in ethylene biosynthesis during fruit ripening. PpyACS8 was highly expressed in the pre-ripening stage and is mostly responsible for the auto-inhibitory low levels of system 1, while PpyACS3 and PpyACS9 were highly expressed after fruit ripening and are associated with the burst of ethylene production in system 2. These results suggest that the type I ACS genes probably obtained unique and redundant functions during evolution.
It has been suggested that plant ACS genes evolved from ACS-like genes, which come from AATase genes [39]. The number and sequence of ACS-like genes are highly conserved in each species identified, which supported the ACS originate evolution of ACS-like genes. ACS-like genes, such as AtACS10 and AtACS12, encode aminotransferases and without ACS activity, they were not generally considered to be authentic ACS genes. However, ACS-like genes in tomato, SlACS11 and SlACS12 display a significant up-regulation during fruit ripening. Li et al. (2015) have reported that ethylene in young apple fruits is mainly produced by catalysis by MdACS6, an ACS-like gene that was mistaken for a type III gene [11,33]. In our data, the expression levels of PpyACS12 and PpyACS13 were low but expressed constitutively throughout the experimental period in both cultivars. Therefore, ACS-like gene expression patterns in tomato, apple and sand pear suggested that ACS-like genes are also responsible for the system 1 ethylene biosynthesis.
Analysis of the pear genome has revealed numerous duplicated genes (paralogs) forming large synteny blocks, covering major portions of chromosomes [27]. Interestingly, all PpyACS genes are distributed on chromosome pairs with partial synteny blocks rather than on chromosome pairs with more than 90% syntenic relations. For instance, PpyACS1 (Ppy15g2510.1) and PpyACS5 (Ppy02g1684.1) were paralogs, and have sequence similarities of more than 94%, located in the chromosome pairs with partial synteny blocks, Chr02 (upper) and Chr15 (middle-upper). The result indicated that PpyACS gene expansion resulted from the WGD event. Interestingly, both PpyACS1 and PpyACS2, two key ACS genes that operate in system1 and system 2 ethylene biosynthesis in fruit ripening, respectively, were located in the middle of chromosome 15. These findings implied that this region is most important in the regulation of fruit ripening and therefore needs to be focused on in future investigations.
The localization of proteins in different organelles may be associated with their function [42,43]. Previous reports have indicated that ACS is localized in the cytoplasm [44][45][46]. Interestingly, subcellular localization predictions show that only type II ACS is located in the cytoplasm and other types of ACS in chloroplasts. Further subcellular localization testing is required to confirm this.

Conclusions
In conclusion, the genome sequencing data showed a total of 13 ACS genes in sand pear, nine of which were novel members. Seven of these ACS genes were found to be involved in sand pear fruit ripening, while four of them were found to be involved in system 1 ethylene biosynthesis, indicating that they play diverse roles in ethylene biosynthesis systems. This study also highlighted the need to include additional ACS genes in studies on ethylene production in sand pear fruits.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/horticulturae7100401/s1, Figure S1: Ripe fruit of 'Ninomiyahakuri' and 'Eli No. 2' pears, Figure S2: Distribution of ACS genes on the Pyrus pyrifolia chromosomes, Figure S3: Species tree of five Rosaceae species and two reference species. Table S1: ACS homologous genes identified in five Rosaceae species and two non-Rosaceae plant species, Table S2: Primers used for qRT-PCR analysis, Table S3: Detailed information of all ACS family genes identified in the seven species, Table S4: Sequences and lengths of motifs among PpyACS proteins, Table S5: ACS orthologs genes identified in the seven species, Table S6